Backport of Python 3.3's 'lzma' module for XZ/LZMA compressed files.
Python 3.3 onwards includes in the standard library module lzma, providing support for working with LZMA and XZ compressed files via the XZ Utils C library (XZ Utils is in a sense LZMA v2). See:
Python’s lzma - http://docs.python.org/dev/library/lzma.html
XZ Utils - http://tukaani.org/xz/
This code is a backport of the Python 3.3 standard library module lzma for use on older versions of Python where it was not included. It is available from PyPI (released downloads only) and GitHub (repository):
There are some older Python libraries like PylibLZMA and PyLZMA but these are both using LZMA Utils (not XZ Utils, so they have no XZ support).
PylibLZMA - http://pypi.python.org/pypi/pyliblzma
PyLZMA - http://pypi.python.org/pypi/pylzma/
LZMA Utils - http://tukaani.org/lzma/
The lmza module provided with Python 3.3 should work on all the main operating systems, so in theory so too should this backport:
Mac OS X: Tested under Python 2.6, 2.7, 3.0 to 3.4 inclusive
Linux: Tested under Python 2.6, 2.7, 3.0 to 3.6 inclusive
Windows: Tested under Python 2.7, 3.6 covering 32-bit and 64-bit, and MSVC and mingw32 compilers
Other than some minor changes in the exceptions for some errors, based on the unit tests everything seems to be working fine.
Support under Python 2.6 and 2.7 appears to be working in that all the appropriate unit tests now pass. Supporting older verions of Python 2 is probably going to be too much work.
We now also support the PyPy implementation of Python 2.7, currently tested with PyPy 5.8.0. It does not currently work on the PyPy implementation of Python 3, but that comes with the lzma standard library module anyway.
I recommend the Conda packaging system which supports Linux, MacOS and Windows. Thanks to the conda-forge package you should be able to install this library with one line, and have the dependencies handled automatically:
$ conda install -c conda-forge backports.lzma
If you are on Linux, there is a good chance that the system packages will include this library and handle the dependencies, e.g. on RedHat/CentOS try:
$ sudo yum install python-backports-lzma
Otherwise, first you must install the XZ Utils C library. On RedHat or CentOS Linux sytems, try:
$ sudo yum install xz-devel
On a Debian based Linux distribution use:
$ sudo apt-get install liblzma-dev
Otherwise do this from source, this is what I do on Mac OS X:
$ curl -L -O http://tukaani.org/xz/xz-5.0.4.tar.gz $ tar -zxvf xz-5.0.4.tar.gz $ cd xz-5.0.4 $ ./configure --prefix=$HOME $ make $ make check $ make install
Now you can install this lzma backport. If using pip, this should work:
$ pip install backports.lzma
Otherwise, you can compile this the old fashioned way. First download and decompress the source code, or clone the github repository:
$ git clone git://github.com/peterjc/backports.lzma.git $ cd backports.lzma $ python setup.py install $ cd test $ python test_lzma.py
To install for a specific version of Python, replace python (which will use the system’s default Python) in the above with a specific version like python2, python2.6 or python3, python3.2, etc.
This should find the XZ Util header file and library automatically (and will check for a local install under your home directory). You should now be able to import the backport from Python as shown below.
If you are trying to install this under the system Python, you will need admin rights and replace python setup.py install with sudo python setup.py install instead.
The expected usage is as follows if you want to prioritise the standard library provided lzma if present:
try: import lzma except ImportError: from backports import lzma #Then use lzma as normal, for example: assert b"Hello!" == lzma.decompress(lzma.compress(b"Hello!"))
Please refer to the lzma documentation online: http://docs.python.org/dev/library/lzma.html
Note that while lzma should be available on Python 3.3, you can still install the backport. This is useful for two reasons, first testing the two act the same way, and second it is possible that your Python installation lacks the standard library lzma. This can happen if Python was installed from source and XZ Utils was not available. If this was a systems level Python install, as a user you could still install XZ Utils and this backport under your own account.
This is using the shared backports namespace introduced by Brandon Rhodes as documented here: http://pypi.python.org/pypi/backports/ and http://bitbucket.org/brandon/backports
- v0.0.1 - January 2013
First public release
- v0.0.2 - April 2013
Fix the seekable attribute on Python 2 (Tomer Chachamu)
More search paths for lib/include headers (Wynn Wilkes)
- v0.0.3 - June 2014
Supports unicode filenames on Python 2 (Irving Reid)
- v0.0.4 - September 2014
Declare namespace package to avoid warnings (Ralph Bean) (Later retracted from PyPI due to installation problems with setuptools versus distutils, see GitHub issue #8 and #9).
- v0.0.5 - June 2016
Backported fix for Python Issue 19839 to ignore non-LZMA trailing data (original Python 3.5.1 patch by Nadeem Vawda, backported by Deroko, see GitHub pull request #5).
- v0.0.6 - June 2016
Updated namespace packaging declaration now required by more recent versions of setuptools which prevented simple installation of v0.0.4 and v0.0.5 from PyPI.
- v0.0.7 - February 2017
Check and prefer the sys.prefix at installation time to find the lib and include headers (John Kirkham).
- v0.0.8 - February 2017
Switch to using README.rst for this document in order to display nicely on PyPI.
- v0.0.9 - 3 January 2018
Now compiles under Windows with passing tests, checked under AppVeyor (see GitHub pull request #25 by Nehal J Wani).
- v0.0.10 - 8 January 2018
Now supports PyPy (specifically their Python 2 implementation, but not yet pypy3 which implements Python 3; see GitHub pull requests #27 and #29 by Michał Górny).
- v0.0.11 - 16 May 2018
Should address namespace issues in v0.0.4, v0.0.5 and v0.0.6 related to a problem in setuptools, and causing side effects with other backports (see pull request #32 from Toshio Kuratomi, and issues #8, #16 and #28).
The initial Python lzma module implementation was by Per Øyvind Karlsen, which was then rewritten by Nadeem Vawda and included with Python 3.3. Based on this work, it was backported to also run on Python 2.6, 2.7 and 3.0, 3.1 and 3.2 by Peter Cock.
Later contributors include: Tomer Chachamu, Wynn Wilkes, Irving Reid, Ralph Bean, Deroko, John Kirkham, Nehal J Wani, Michał Górny, Toshio Kuratomi.
Please report any reproducible bugs via the GitHub issue tracker at https://github.com/peterjc/backports.lzma/issues including details about your operating system, version of Python, XY Utils, the lzma backport etc. Reproducible test cases are particularly helpful.
If you can demonstrate a problem in this backport but not in the lzma module included with Python 3.3 or later, then it is clearly something we will need to fix.
Any issues in the lzma module as bundled with Python 3.3 or later should be reported to the Python project at http://bugs.python.org instead (and we can hopefully apply any official fix to the backport as well).
The version is incremented in file backports/lzma/__init__.py (from where setup.py will extract it at runtime).
After testing locally and with TravisCI (see below), new releases are tagged in git as follows:
$ git tag backports.lzma.vX.X.X
Tags must explicitly be pushed to GitHub:
$ git push origin master --tags
I then use the following to upload a new release to the Python Packaging Index (PyPI):
$ python setup.py sdist $ twine upload dist/backports.lzma-X.X.X.tar.gz
If not already installed, try pip install twine.
The update then appears on http://pypi.python.org/pypi/backports.lzma/
TravisCI is being used for continuous integration testing under Linux, see https://travis-ci.org/peterjc/backports.lzma
Similarly, AppVeyor is being used for testing under Windows, see: https://ci.appveyor.com/project/peterjc/backports-lzma/history
Release history Release notifications | RSS feed
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.